Ultrasonic image gain compensation method, device, equipment and medium

Abstract

This application relates to an ultrasound image gain compensation method, apparatus, device, and medium. The method comprises acquiring a first ultrasound image to be compensated; inputting the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; and performing gain compensation on the first ultrasound image to be compensated based on the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image. The ultrasound image gain compensation method provided in this application eliminates the need for manual user adjustment, thereby reducing user workload and working time.

Description

Technical Field

[0001] The present application relates to the field of ultrasonic detection technology, and in particular to an ultrasonic image gain compensation method, device, equipment and medium. Background Art

[0002] Medical ultrasound detection technology has been widely used in modern medical diagnosis and treatment. During the ultrasound imaging process using ultrasound detection technology, it is necessary to adjust the gain to compensate the ultrasound image.

[0003] Gain adjustment includes adjusting time gain compensation (TGC) or lateral gain compensation (LGC). In conventional technology, ultrasound images are typically gain compensated according to a preset time gain compensation curve and / or a preset lateral gain compensation curve.

[0004] However, ultrasonic detection technology has many application scenarios, and the compensation effect of the preset gain compensation curve is poor. The user needs to make manual adjustments later, which increases the user's workload and working time. Summary of the Invention

[0005] Based on this, it is necessary to provide an ultrasound image gain compensation method, device, equipment and medium to address the above technical problems.

[0006] In a first aspect, an embodiment of the present application provides an ultrasound image gain compensation method, the method comprising:

[0007] acquiring a first ultrasound image to be compensated;

[0008] Inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0009] Gain compensation is performed on the first ultrasonic image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image.

[0010] In one embodiment, the gain compensation model includes a first model and a second model. The first ultrasound image to be compensated is input into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve, including:

[0011] Inputting a first ultrasound image to be compensated into a first model to obtain a time gain compensation curve;

[0012] The first ultrasound image to be compensated is input into the second model to obtain a lateral gain compensation curve.

[0013] In one embodiment, the gain compensation model includes a convolutional residual module, a pooling layer, and a fully connected layer. The first ultrasound image to be compensated is input into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve, including:

[0014] Inputting the first ultrasound image to be compensated into the convolution residual module, performing feature extraction on the first ultrasound image to be compensated by the convolution residual module to obtain image feature data;

[0015] The image feature data is input into the pooling layer, and the image feature data is subjected to dimensionality reduction processing by the pooling layer to obtain reduced dimensionality data;

[0016] The dimension-reduced data is input into the fully connected layer to obtain the time gain compensation curve and the lateral gain compensation curve.

[0017] In one embodiment, the gain compensation model includes a plurality of convolutional residual modules, and the plurality of convolutional residual modules are jump-connected.

[0018] In one embodiment, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0019] In one embodiment, a first ultrasound image to be compensated is input into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; and gain compensation is performed on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image, including:

[0020] Inputting the first ultrasound image to be compensated into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model obtained by updating the parameters of the gain compensation model;

[0021] Gain compensation is performed on the first ultrasonic image to be compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasonic image.

[0022] In one embodiment, the training method of the third model includes:

[0023] Acquire an ultrasound image sample to be compensated, input the ultrasound image sample to be compensated into a gain compensation model, and obtain a time gain compensation curve sample and a lateral gain compensation curve sample;

[0024] Acquire a sample of a time gain compensation curve after fine-tuning and a sample of a lateral gain compensation curve after fine-tuning;

[0025] The gain compensation model is trained based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0026] In a second aspect, an embodiment of the present application provides an ultrasound image gain compensation device, the device comprising:

[0027] an acquisition module, configured to acquire a first ultrasound image to be compensated;

[0028] a determination module, configured to input the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0029] The compensation module is configured to perform gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image.

[0030] In a third aspect, an embodiment of the present application provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the method provided in the above embodiment when executing the computer program.

[0031] In a fourth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon, which implements the steps of the method provided in the above embodiment when the computer program is executed by a processor.

[0032] In a fifth aspect, an embodiment of the present application further provides a computer program product, comprising a computer program, which implements the steps of the method provided in the above embodiment when the computer program is executed by a processor.

[0033] The embodiments of the present application provide a method, device, equipment and medium for gain compensation of ultrasonic images. The method obtains a time gain compensation curve and a lateral gain compensation curve by inputting the acquired first ultrasonic image to be compensated into a gain compensation model; the first ultrasonic image to be compensated is gain compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image. The ultrasonic image gain compensation method provided in this embodiment uses a pre-trained gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve that can be used to gain compensate the first ultrasonic image to be compensated. The user does not need to manually adjust the first ultrasonic image to be compensated, thereby reducing the user's workload and working time. At the same time, this embodiment can simultaneously perform time gain compensation and lateral gain compensation on the first ultrasonic image to be compensated, which can improve the efficiency of gain compensation of the first ultrasonic image to be compensated, and can also improve the effect of gain compensation of the first ultrasonic image to be compensated. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the conventional technology, the following briefly introduces the drawings required for use in the embodiments or the conventional technology descriptions. Obviously, the drawings described below are only some embodiments of the present application. For different technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0035] Figure 1 A diagram illustrating an application scenario of an ultrasound image gain compensation method provided by an embodiment;

[0036] Figure 2 A schematic flow chart of the steps of an ultrasound image gain compensation method provided in one embodiment;

[0037] Figure 3 A schematic flow chart of the steps of an ultrasound image gain compensation method provided in another embodiment;

[0038] Figure 4 A schematic diagram of the structure of a gain compensation model provided by an embodiment;

[0039] Figure 5 A schematic flow chart of the steps of an ultrasound image gain compensation method provided in another embodiment;

[0040] Figure 6 A schematic structural diagram of a gain compensation model provided in another embodiment;

[0041] Figure 7 A schematic flow chart of the steps of an ultrasound image gain compensation method provided in another embodiment;

[0042] Figure 8 A schematic flow chart of the steps of an ultrasound image gain compensation method provided in another embodiment;

[0043] Figure 9 A schematic structural diagram of an ultrasonic image gain compensation device provided in one embodiment;

[0044] Figure 10 A schematic diagram of the structure of a computer device provided in one embodiment. DETAILED DESCRIPTION

[0045] To make the above-mentioned objects, features, and advantages of the present application more clearly understood, the specific embodiments of the present application are described in detail below with reference to the accompanying drawings. The following description sets forth many specific details to facilitate a full understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar improvements without violating the scope of the present application. Therefore, the present application is not limited to the specific embodiments disclosed below.

[0046] The serial numbers assigned to the components in this document, such as "first", "second", etc., are only used to distinguish the objects described and do not have any order or technical meaning.

[0047] The ultrasound image gain compensation method provided in the embodiment of the present application can be applied to Figure 1 In the application scenario shown, the application environment includes a terminal 100 and an ultrasonic imaging device 200, wherein the terminal 100 can communicate with the ultrasonic imaging device 200 via a network. The terminal 100 can be, but is not limited to, various personal computers, laptops, and tablet computers. This embodiment does not limit the specific structure of the ultrasonic imaging device 200.

[0048] Traditionally, a terminal acquires a noise image from an unscanned ultrasound scan to obtain noise image data for the image to be processed. Based on the obtained noise image data, the processed image data is denoised. The type of the processed image data after denoising is determined. If the processed image data after denoising is tissue image data, the denoised tissue image is segmented to extract the tissue regions within the image. Based on the extracted tissue regions, the overall gain is calculated and applied to the overall image. This method allows the terminal to adapt gain adjustment to different application scenarios, but it cannot accurately segment tissue or determine the type of segmented tissue, making it difficult to achieve optimal gain improvement.

[0049] Conventional technology involves detecting a user's touch signal applied to an ultrasound image and determining a region of interest (ROI) on the image based on the touch signal. Time gain compensation and lateral gain compensation corresponding to the ROI are then determined based on the touch signal. The ROI is then adjusted based on the time and lateral gain compensations, and the adjusted ultrasound image is displayed. This method allows simultaneous adjustment of both time and lateral gain, but it requires manual operation and potentially requires user experience.

[0050] To this end, the present application provides an ultrasound image gain compensation method.

[0051] The following specific embodiments describe in detail the technical solution of this application and how the technical solution of this application solves the technical problem. The following specific embodiments can be combined with each other, and the same or similar concepts or processes may not be repeated in some embodiments. The embodiments of this application will be described below in conjunction with the accompanying drawings.

[0052] See Figure 2 , an embodiment of the present application provides an ultrasound image gain compensation method. This embodiment applies this method to Figure 1 For illustration purposes, the following steps are involved:

[0053] Step 200: Acquire a first ultrasound image to be compensated.

[0054] The first ultrasound image to be compensated may be a Figure 1 The ultrasonic imaging device in the apparatus obtains an ultrasonic image whose brightness needs to be adjusted. The first ultrasonic image to be compensated can be stored in a memory of the terminal, and the terminal can directly call it from the memory when needed. The first ultrasonic image to be compensated can also be stored in a storage unit of the ultrasonic imaging device, and the terminal can retrieve it from the storage unit of the ultrasonic imaging device when needed. This embodiment does not limit the specific method of obtaining the first ultrasonic image to be compensated, as long as its function can be achieved.

[0055] Step 210: Input the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve.

[0056] Time gain compensation (TGC) is a technique used by ultrasound imaging equipment to compensate for the depth of the image (the direction of sound wave transmission). As sound waves propagate through different tissues, their intensity attenuates, causing the received echo to appear stronger near-field and weaker far-field. This results in uneven brightness in the ultrasound image across depth. Time-based TGC can achieve uniform brightness across depth.

[0057] Lateral gain compensation is applied to ultrasonic imaging devices in the element direction (element channel). When the sound beam is centered on the probe, the most contributing elements contribute, and the beam echo is the strongest. When the beam is on either side of the probe, the contributing elements gradually decrease, and the beam echo weakens. Gain compensation is applied based on the element channel, achieving a balanced brightness of the ultrasound image along the element channel direction.

[0058] After obtaining the first ultrasound image to be compensated, the terminal inputs it into a pre-trained gain compensation model. The gain compensation model processes the image and can output a time gain compensation curve and a lateral gain compensation curve. In other words, the pre-trained gain compensation model can include an input channel and two output channels. The first ultrasound image to be compensated is input into the gain compensation model through the input channel. After the gain compensation model processes the image, the time gain compensation curve and the lateral gain compensation curve are output through the two output channels. The gain compensation model is an independent model. This embodiment does not limit the training method of the gain compensation model or the specific structure of the gain compensation model, as long as its function can be achieved.

[0059] In an optional embodiment, the gain compensation model can be obtained by pre-training the initial neural network by the terminal according to the first training sample. The first training sample includes the ultrasound image sample to be compensated and the gain compensation curve and lateral gain compensation curve corresponding to each ultrasound image in the ultrasound image sample to be compensated.

[0060] Step 220 : Perform gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image.

[0061] After obtaining the time gain compensation curve and the lateral gain compensation curve, the terminal uses the time gain compensation curve to adjust the brightness of the first ultrasound image to be compensated in the depth direction, and uses the lateral gain compensation curve to adjust the brightness of the first ultrasound image to be compensated in the array element channel direction to obtain an adjusted target ultrasound image.

[0062] The ultrasonic image gain compensation method provided in the embodiment of the present application is as follows: obtaining a first ultrasonic image to be compensated; inputting the first ultrasonic image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; and performing gain compensation on the first ultrasonic image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image. The ultrasonic image gain compensation method provided in the present embodiment uses a pre-trained gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve that can be used to gain compensate the first ultrasonic image to be compensated, without the user having to manually adjust the first ultrasonic image to be compensated, thereby reducing the user's workload and working time. At the same time, the present embodiment can simultaneously perform time gain compensation and lateral gain compensation on the first ultrasonic image to be compensated, thereby improving the efficiency of gain compensation for the first ultrasonic image to be compensated, and also improving the effect of gain compensation for the first ultrasonic image to be compensated.

[0063] See Figure 3 In one embodiment, the gain compensation model includes a first model and a second model. In other words, the gain compensation model can be two independent models: the first model and the second model, and the first model and the second model are trained separately.

[0064] In an optional embodiment, the first model may be obtained by pre-training a neural network model by the terminal based on a second training sample. The second training sample includes an ultrasound image sample to be compensated and a time gain compensation curve corresponding to each ultrasound image in the ultrasound image sample to be compensated. The second model may be obtained by pre-training a neural network model by the terminal based on a third training sample. The third training sample includes an ultrasound image sample to be compensated and a lateral gain compensation curve corresponding to each ultrasound image in the ultrasound image sample to be compensated.

[0065] When the gain compensation model includes a first model and a second model, a possible implementation method of inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve includes the following steps:

[0066] Step 300: Input a first ultrasound image to be compensated into a first model to obtain a time gain compensation curve.

[0067] After obtaining the first ultrasound image to be compensated, the terminal inputs it into the first model. The first model then processes the image and outputs a time gain compensation curve corresponding to the first ultrasound image to be compensated. In other words, the first model includes an input channel and an output channel. The terminal inputs the first ultrasound image to be compensated into the first model through the input channel, and the output channel of the first model outputs the time gain compensation curve corresponding to the first ultrasound image to be compensated.

[0068] Step 310: Input the first ultrasound image to be compensated into the second model to obtain a lateral gain compensation curve.

[0069] After obtaining the first ultrasound image to be compensated, the terminal inputs it into the second model. The second model then processes the image and outputs a lateral gain compensation curve corresponding to the first ultrasound image to be compensated. In other words, the second model includes an input channel and an output channel. The terminal inputs the first ultrasound image to be compensated into the second model through the input channel, and the output channel of the second model outputs the lateral gain compensation curve corresponding to the first ultrasound image to be compensated.

[0070] In this embodiment, the gain compensation model includes a first model and a second model. Thus, when using the gain compensation model to determine a time gain compensation curve and a lateral gain compensation curve corresponding to a first ultrasound image to be compensated, the time gain compensation curve and the lateral gain compensation curve are determined using two independent first and second models, respectively, thereby improving determination efficiency and accuracy. Furthermore, when training the gain compensation model, training the first and second models separately can improve model training efficiency and accuracy.

[0071] In one embodiment, the structure of the gain compensation model is as follows: Figure 4 As shown, the gain compensation model includes a convolutional residual module, a pooling layer, and a fully connected layer. The convolutional residual module, the pooling layer, and the fully connected layer are connected in sequence. The input of the convolutional residual module serves as the input of the gain compensation model. That is, the first ultrasound image to be compensated is input from the input of the convolutional residual module to the convolutional residual module. The output of the fully connected layer serves as the output of the gain compensation model. That is, the time gain compensation curve and the lateral gain compensation curve are output from the output of the fully connected layer.

[0072] A possible implementation method involves inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve. Figure 5 Shown include:

[0073] Step 500: Input a first ultrasound image to be compensated into a convolution residual module, and perform feature extraction on the first ultrasound image to be compensated by the convolution residual module to obtain image feature data.

[0074] After the terminal inputs the first ultrasound image to be compensated into the gain compensation model, it first passes it through the convolution residual module within the gain compensation model. This convolution residual module extracts features from the first ultrasound image to generate image feature data. This image feature data is used to construct the time gain compensation curve and lateral gain compensation curve output by the gain compensation model.

[0075] Step 510: Input the image feature data into the pooling layer, and perform dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data.

[0076] The image feature data obtained by the terminal through the convolution residual module is multi-dimensional data. After the terminal inputs the image feature data output from the convolution residual module into the pooling layer, the pooling layer will perform dimensionality reduction processing on the multi-dimensional image feature data to obtain one-dimensional reduced dimensionality data.

[0077] Step 520: Input the dimension-reduced data into the fully connected layer to obtain a time gain compensation curve and a lateral gain compensation curve.

[0078] The terminal inputs the dimensionality reduction data output from the pooling layer into the fully connected layer, and performs comprehensive processing on the dimensionality reduction data through the fully connected layer, that is, comprehensively processes the characteristics extracted from the first ultrasound image to be compensated, and outputs the time gain compensation curve and the lateral gain compensation curve from the fully connected layer.

[0079] This embodiment uses a convolutional residual module, a pooling layer, and a fully connected layer to perform feature extraction, dimensionality reduction, and comprehensive processing on the first ultrasound image to be compensated, thereby obtaining accurate time gain compensation curves and lateral gain compensation curves. This embodiment does not limit the specific structure of the convolutional residual module, the pooling layer, and the fully connected layer, as long as their functions can be achieved.

[0080] In a specific embodiment, the image feature data output by the convolutional residual module is three-dimensional data K*N*M. This data is then processed through a pooling layer to reduce its dimensionality, yielding K*1*1 dimensional data. This data is then processed through a fully connected layer to produce two-dimensional data 2*L. Here, 2 represents the output of two curves (a time gain compensation curve and a lateral gain compensation curve) through the fully connected layer, and L represents the number of sampling points for each curve. The time gain compensation curve and the lateral gain compensation curve can be generated based on the sampling points of each curve.

[0081] In an optional embodiment, the convolutional residual module can be replaced by multiple modules, such as a DPN (Dual Path Network) module.

[0082] In an optional embodiment, when the gain compensation model includes a first model and a second model, the first model may include a convolutional residual module, a pooling layer, and a fully connected layer, and the second model also includes a convolutional residual module, a pooling layer, and a fully connected layer.

[0083] In one embodiment, the structure of the gain compensation model is as follows: Figure 6 As shown, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0084] The convolutional layer extracts features from the first ultrasound image to be compensated. The batch normalization layer normalizes the extracted feature data, ensuring that it conforms to a normal distribution with a mean of 0 and a variance of 1. The batch normalization layer mitigates vanishing and exploding gradients during gain compensation model training, improving the training rate. The activation layer incorporates nonlinear factors, addressing problems that linear models cannot.

[0085] In one embodiment, the gain compensation model includes multiple convolutional residual modules, and the multiple convolutional residual modules are jump-connected.

[0086] The deeper the neural network corresponding to the gain compensation model, the more information it extracts and the richer the features. Multiple convolutional residual modules can solve the problems of gradient explosion and gradient disappearance caused by the deepening of the neural network through jump connections, so that the neural network can be deepened to a deeper level, extract more information, and supplement the lost data, thereby improving the accuracy of the trained gain compensation model. For example, the gain compensation model includes 3 convolutional residual modules, and the jump connections of the 3 convolutional residual modules are as follows: Figure 6 As shown, the input of the two sequentially connected convolution residual modules is the first ultrasound image to be compensated, and the input of the other convolution residual module is the first ultrasound image to be compensated. The outputs of the two sequentially connected convolution residual modules and the output of the other convolution residual module are summed to obtain image feature data.

[0087] See Figure 7 In one embodiment, a possible implementation method involves inputting a first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; performing gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image, including the following steps:

[0088] Step 700: Input the first ultrasound image to be compensated into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model obtained by updating the parameters of the gain compensation model.

[0089] The third model is obtained by pre-training the gain compensation model (updating the gain compensation model's parameters). The third model includes one input channel and two output channels. After obtaining the third model, the terminal inputs the first ultrasound image to be compensated into the third model. The third model's two output channels then output a new time gain compensation curve and a new lateral gain compensation curve.

[0090] In one embodiment, the steps of the training method of the third model are as follows Figure 8 Shown, including:

[0091] Step 800: Acquire ultrasound image samples to be compensated, input the ultrasound image samples to be compensated into a gain compensation model, and obtain time gain compensation curve samples and lateral gain compensation curve samples.

[0092] For a description of the gain compensation model, reference may be made to the specific description of the above embodiment and will not be repeated here. The ultrasound image samples to be compensated include multiple ultrasound images to be compensated. Each ultrasound image to be compensated is input into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve corresponding to the ultrasound image to be compensated, thereby obtaining time gain compensation curve samples and lateral gain compensation curve samples corresponding to the ultrasound image samples to be compensated.

[0093] Step 810: Acquire a sample of a fine-tuned time gain compensation curve and a sample of a fine-tuned lateral gain compensation curve.

[0094] The fine-tuned time gain compensation curve samples are obtained by fine-tuning each time gain compensation curve in the time gain compensation curve samples; the fine-tuned lateral gain compensation curve samples are obtained by fine-tuning each lateral gain compensation curve in the lateral gain compensation curve samples. The fine-tuned time gain compensation curve samples and the fine-tuned lateral gain compensation curve samples are in accordance with user preferences. This embodiment does not limit the specific method for obtaining the fine-tuned time gain compensation curve samples and the fine-tuned lateral gain compensation curve samples, as long as the method can achieve the desired effect.

[0095] In an optional embodiment, the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample are obtained by a user through fine-tuning and stored in a memory of the terminal. Specifically, the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample can be obtained by the user adjusting the time gain compensation curve sample and the lateral gain compensation curve sample according to the user's requirements for ultrasound image brightness distribution.

[0096] In another optional embodiment, the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample are obtained by, after obtaining the time gain compensation curve sample and the lateral gain compensation curve sample, fine-tuning the time gain compensation curve sample and the lateral gain compensation curve sample, respectively, using a preset algorithm or a preset fine-tuning model. Specifically, the preset algorithm or the preset fine-tuning model may be an algorithm preset by a user based on user requirements for ultrasound image brightness distribution, or a pre-trained model.

[0097] Step 820 : Train the gain compensation model based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0098] After obtaining the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample, the terminal uses the ultrasound image sample to be compensated, the fine-tuned time gain compensation curve sample, and the fine-tuned lateral gain compensation curve sample to perform supervised training on the gain compensation model, that is, the parameters in the gain compensation model are updated to obtain a third model.

[0099] Step 710 : Perform gain compensation on the first ultrasound image to be compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasound image.

[0100] The terminal inputs the first ultrasound image to be compensated into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve. Gain compensation is performed on the first ultrasound image to be compensated based on the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasound image. Because the target ultrasound image is obtained by compensating based on the new time gain compensation curve and the new lateral gain compensation curve, the target ultrasound image is different from an ultrasound image obtained using the gain compensation model.

[0101] In this embodiment, the third model is trained based on time gain compensation curve samples and lateral gain compensation curve samples that meet the user's preferences. The first ultrasound image to be compensated is gain compensated based on the new time gain compensation curve and the new lateral gain compensation curve obtained by the third model, and a target ultrasound image that meets the user's preferences can be obtained, which can improve the applicability of the ultrasound image gain compensation method.

[0102] It should be understood that, although the steps in the flowcharts of the above-mentioned embodiments are shown in sequence as indicated by the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and these steps can be executed in other orders. Moreover, at least a portion of the steps in the flowcharts of the above-mentioned embodiments may include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a portion of steps or stages in other steps.

[0103] Based on the same inventive concept, embodiments of the present application also provide an ultrasonic image gain compensation device for implementing the aforementioned ultrasonic image gain compensation method. The solution provided by this device is similar to the solution described in the aforementioned method. Therefore, the specific limitations in one or more of the following embodiments of the ultrasonic image gain compensation device can be found in the aforementioned limitations of the ultrasonic image gain compensation method and will not be further elaborated here.

[0104] See Figure 9 An embodiment of the present application provides an ultrasound image gain compensation device 10, which includes an acquisition module 11, a determination module 12 and a compensation module 13.

[0105] The acquisition module 11 is used to acquire a first ultrasound image to be compensated;

[0106] The determination module 12 is configured to input the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0107] The compensation module 13 is configured to perform gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image.

[0108] In one embodiment, the determination module 12 includes a first determination unit and a second determination unit. The first determination unit is configured to input the first ultrasound image to be compensated into a first model to obtain a time gain compensation curve; and the second determination unit is configured to input the first ultrasound image to be compensated into a second model to obtain a lateral gain compensation curve.

[0109] In one embodiment, the determination module 12 is specifically used to input the first ultrasound image to be compensated into the convolution residual module, perform feature extraction on the first ultrasound image to be compensated through the convolution residual module to obtain image feature data; input the image feature data into the pooling layer, perform dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data; input the reduced dimensionality data into the fully connected layer to obtain a time gain compensation curve and a lateral gain compensation curve.

[0110] In one embodiment, the gain compensation model includes multiple convolutional residual modules, and the multiple convolutional residual modules are jump-connected.

[0111] In one embodiment, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0112] In one embodiment, the target ultrasonic image determination module is used to input the first ultrasonic image to be compensated into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model after the parameters of the gain compensation model are updated; the first ultrasonic image to be compensated is gain compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasonic image.

[0113] In one embodiment, the ultrasonic image gain compensation device 10 also includes a third model training module, which is used to obtain ultrasonic image samples to be compensated, input the ultrasonic image samples to be compensated into the gain compensation model, and obtain time gain compensation curve samples and lateral gain compensation curve samples; obtain fine-tuned time gain compensation curve samples and fine-tuned lateral gain compensation curve samples; train the gain compensation model based on the ultrasonic image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0114] Each module in the ultrasound image gain compensation device can be implemented in whole or in part through software, hardware, or a combination thereof. Each module can be embedded in or independent of a processor in a computer device in hardware form, or can be stored in a computer device memory in software form, so that the processor can call and execute the corresponding operations of each module.

[0115] In one embodiment, a computer device is provided. The computer device may be a terminal, and its internal structure diagram may be as follows: Figure 10As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, an ultrasonic image gain compensation method is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covering the display screen, or a button, trackball or touchpad provided on the computer device housing, or an external keyboard, touchpad or mouse.

[0116] Those skilled in the art will understand that Figure 10 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than shown in the figure, or combine certain components, or have a different component arrangement.

[0117] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the following steps are performed:

[0118] acquiring a first ultrasound image to be compensated;

[0119] Inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0120] Gain compensation is performed on the first ultrasonic image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image.

[0121] In one embodiment, when the processor executes the computer program, the following steps are further implemented: inputting the first ultrasound image to be compensated into the first model to obtain a time gain compensation curve; inputting the first ultrasound image to be compensated into the second model to obtain a lateral gain compensation curve.

[0122] In one embodiment, when the processor executes the computer program, the following steps are further implemented: inputting the first ultrasound image to be compensated into the convolution residual module, performing feature extraction on the first ultrasound image to be compensated through the convolution residual module to obtain image feature data; inputting the image feature data into the pooling layer, performing dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data; inputting the reduced dimensionality data into the fully connected layer to obtain a time gain compensation curve and a lateral gain compensation curve.

[0123] In one embodiment, the gain compensation model includes multiple convolutional residual modules, and the multiple convolutional residual modules are jump-connected.

[0124] In one embodiment, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0125] In one embodiment, when the processor executes the computer program, the following steps are also implemented: the first ultrasound image to be compensated is input into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model after the parameters of the gain compensation model are updated; and the first ultrasound image to be compensated is gain compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasound image.

[0126] In one embodiment, when the processor executes the computer program, the following steps are further implemented: obtaining ultrasound image samples to be compensated, inputting the ultrasound image samples to be compensated into a gain compensation model to obtain time gain compensation curve samples and lateral gain compensation curve samples; obtaining fine-tuned time gain compensation curve samples and fine-tuned lateral gain compensation curve samples; and training the gain compensation model based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0127] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented:

[0128] acquiring a first ultrasound image to be compensated;

[0129] Inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0130] Gain compensation is performed on the first ultrasonic image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image.

[0131] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: inputting the first ultrasound image to be compensated into the first model to obtain a time gain compensation curve; inputting the first ultrasound image to be compensated into the second model to obtain a lateral gain compensation curve.

[0132] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: inputting the first ultrasound image to be compensated into the convolution residual module, performing feature extraction on the first ultrasound image to be compensated through the convolution residual module to obtain image feature data; inputting the image feature data into the pooling layer, performing dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data; inputting the reduced dimensionality data into the fully connected layer to obtain a time gain compensation curve and a lateral gain compensation curve.

[0133] In one embodiment, the gain compensation model includes multiple convolutional residual modules, and the multiple convolutional residual modules are jump-connected.

[0134] In one embodiment, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0135] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: the first ultrasound image to be compensated is input into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model after the parameters of the gain compensation model are updated; the first ultrasound image to be compensated is gain compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasound image.

[0136] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining ultrasound image samples to be compensated, inputting the ultrasound image samples to be compensated into the gain compensation model to obtain time gain compensation curve samples and lateral gain compensation curve samples; obtaining fine-tuned time gain compensation curve samples and fine-tuned lateral gain compensation curve samples; training the gain compensation model based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0137] In one embodiment, a computer program product is provided, comprising a computer program, which, when executed by a processor, implements the following steps:

[0138] acquiring a first ultrasound image to be compensated;

[0139] Inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve;

[0140] Gain compensation is performed on the first ultrasonic image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasonic image.

[0141] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: inputting the first ultrasound image to be compensated into the first model to obtain a time gain compensation curve; inputting the first ultrasound image to be compensated into the second model to obtain a lateral gain compensation curve.

[0142] In one embodiment, when the computer program is executed by the processor, the following steps are further implemented: inputting the first ultrasound image to be compensated into the convolution residual module, performing feature extraction on the first ultrasound image to be compensated through the convolution residual module to obtain image feature data; inputting the image feature data into the pooling layer, performing dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data; inputting the reduced dimensionality data into the fully connected layer to obtain a time gain compensation curve and a lateral gain compensation curve.

[0143] In one embodiment, the gain compensation model includes multiple convolutional residual modules, and the multiple convolutional residual modules are jump-connected.

[0144] In one embodiment, the convolutional residual module includes a batch normalization layer, an activation layer, and at least one convolutional layer.

[0145] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: the first ultrasound image to be compensated is input into the third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model after the parameters of the gain compensation model are updated; the first ultrasound image to be compensated is gain compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain a target ultrasound image.

[0146] In one embodiment, when the computer program is executed by the processor, the following steps are also implemented: obtaining ultrasound image samples to be compensated, inputting the ultrasound image samples to be compensated into the gain compensation model to obtain time gain compensation curve samples and lateral gain compensation curve samples; obtaining fine-tuned time gain compensation curve samples and fine-tuned lateral gain compensation curve samples; training the gain compensation model based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain a third model.

[0147] Those skilled in the art will appreciate that all or part of the processes in the above-mentioned embodiment methods can be implemented by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, database or other media used in the embodiments provided in this application may include at least one of non-volatile and volatile memory. Non-volatile memory may include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory may include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in the various embodiments provided herein may include at least one of a relational database and a non-relational database. Non-relational databases may include, but are not limited to, distributed databases based on blockchains. The processor involved in the various embodiments provided herein may be, but are not limited to, a general-purpose processor, a central processing unit, a graphics processing unit, a digital signal processor, a programmable logic unit, a data processing logic unit based on quantum computing, and the like.

[0148] The technical features of the above embodiments can be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0149] The above embodiments merely illustrate several implementation methods of the present application. While the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the present application. It should be noted that a person skilled in the art could make various modifications and improvements without departing from the spirit of the present application, all of which fall within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be determined by the appended claims.

Claims

1. A method for ultrasonic image gain compensation, characterized in that: The method comprises: acquiring a first ultrasound image to be compensated; Inputting the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; performing gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image; The step of inputting the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; and performing gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image includes: Inputting the first ultrasound image to be compensated into a third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model obtained by updating the parameters of the gain compensation model using the ultrasound image sample to be compensated, the fine-tuned time gain compensation curve sample, and the fine-tuned lateral gain compensation curve sample, and the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample are obtained by fine-tuning the time gain compensation curve sample and the lateral gain compensation curve sample using a preset algorithm or a preset fine-tuning model; Gain compensation is performed on the first ultrasound image to be compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain the target ultrasound image.

2. The ultrasonic image gain compensation method according to claim 1, characterized in that: The gain compensation model includes a first model and a second model, and inputting the first ultrasound image to be compensated into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve includes: Inputting the first ultrasound image to be compensated into the first model to obtain the time gain compensation curve; The first ultrasound image to be compensated is input into the second model to obtain the lateral gain compensation curve.

3. The ultrasonic image gain compensation method according to claim 1, wherein: The gain compensation model includes a convolution residual module, a pooling layer, and a fully connected layer. The first ultrasound image to be compensated is input into the gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve, including: Inputting the first ultrasound image to be compensated into the convolution residual module, and performing feature extraction on the first ultrasound image to be compensated by the convolution residual module to obtain image feature data; Inputting the image feature data into the pooling layer, and performing dimensionality reduction processing on the image feature data through the pooling layer to obtain reduced dimensionality data; The dimension-reduced data is input into the fully connected layer to obtain the time gain compensation curve and the lateral gain compensation curve.

4. The ultrasonic image gain compensation method according to claim 3, characterized in that: The gain compensation model includes a plurality of convolutional residual modules, and the plurality of convolutional residual modules are jump-connected.

5. The ultrasonic image gain compensation method according to claim 4, characterized in that: The gain compensation model includes three convolution residual modules, the inputs of two convolution residual modules connected in sequence are the first ultrasound image to be compensated, and the input of another convolution residual module is the first ultrasound image to be compensated; the outputs of the two convolution residual modules connected in sequence are summed with the output of another convolution residual module to obtain image feature data.

6. The ultrasonic image gain compensation method according to claim 3 or 4, characterized in that: The convolutional residual module includes a batch normalization layer, an activation layer and at least one convolutional layer.

7. The ultrasonic image gain compensation method according to claim 1, characterized in that: The training method of the third model includes: Acquire an ultrasound image sample to be compensated, input the ultrasound image sample to be compensated into the gain compensation model, and obtain a time gain compensation curve sample and a lateral gain compensation curve sample; Acquire a sample of a time gain compensation curve after fine-tuning and a sample of a lateral gain compensation curve after fine-tuning; The gain compensation model is trained based on the ultrasound image samples to be compensated, the fine-tuned time gain compensation curve samples, and the fine-tuned lateral gain compensation curve samples to obtain the third model.

8. An ultrasonic image gain compensation device, characterized in that: The device comprises: an acquisition module, configured to acquire a first ultrasound image to be compensated; a determination module, configured to input the first ultrasound image to be compensated into a gain compensation model to obtain a time gain compensation curve and a lateral gain compensation curve; a compensation module, configured to perform gain compensation on the first ultrasound image to be compensated according to the time gain compensation curve and the lateral gain compensation curve to obtain a target ultrasound image; The compensation module is specifically used to input the first ultrasound image to be compensated into a third model to obtain a new time gain compensation curve and a new lateral gain compensation curve; the third model is a model obtained by updating the parameters of the gain compensation model using the ultrasound image sample to be compensated, the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample, and the fine-tuned time gain compensation curve sample and the fine-tuned lateral gain compensation curve sample are obtained by fine-tuning the time gain compensation curve sample and the lateral gain compensation curve sample using a preset algorithm or a preset fine-tuning model; the first ultrasound image to be compensated is gain compensated according to the new time gain compensation curve and the new lateral gain compensation curve to obtain the target ultrasound image.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images